Our current research is focused on the neuronal networks in the basal ganglia which control eye movements. We have shown that sensory and cognitive signals carried by neurons in the basal ganglia are strongly modified by expected reward. The signals are then transmitted to a midbrain structure called the superior colliculus (SC). The SC is a key structure for sensorimotor transformation. Spatial information contained in sensory signals is converted into orienting movements of the eyes and the head. The sensorimotor conversion is competitive because incoming sensory signals are abundant and complex while orienting movements are made one at a time. Successful behavior, then, requires an efficient selection mechanism. An object or position that was previously associated with reward is more likely to be selected, and consequently the animal is more likely to orient to the position of the object. However, it has been unknown whether and how the SC is involved in the reward-based selection. To answer this question, we recorded from single cell activity from the monkey SC while the subjects (macaque monkeys) performed a memory-guided saccade task with an asymmetric reward schedule. In the memory-guided saccade task, a visual cue stimulus indicated the goal of an upcoming saccade. It also indicated whether or not reward would be given after the saccade. Many SC neurons responded to the visual cue. We found that the visual responses frequently increased when the visual cue indicated an upcoming reward. The increase occurred in two distinct manners: 1) reactively as an increase in the gain of the visual response when the stimulus indicated an upcoming reward; 2) proactively as an increase in anticipatory activity when reward was expected in the neuron?s response field. These effects were observed mostly in saccade-related SC neurons in the deeper layer which would receive inputs from the cortical eye fields and the basal ganglia. We concluded that the SC plays an important role in the reward-based selection of orienting behavior. These results suggest that the reward-based selection in SC neurons is derived mainly from the basal ganglia. Previous studies from our research group showed that neurons in the basal ganglia - caudate nucleus (CD) and the substantia nigra pars reticulata (SNr) - exhibit pre-cue anticipatory changes in activity in the same task. We have shown previously that saccades are controlled by the serial inhibitory connections from the CD to the SNr and from the SNr to the SC. The pre-cue activity in CD neurons would inhibit the rapid firing of SNr neurons, and therefore SC neurons would be released from the tonic inhibition by the SNr. Our demonstration of the bias effect in SC neurons supplies the crucial missing link between the neuronal activity in the basal ganglia and eye movements. This neuron-behavior relation can be used as an excellent system in which pathological mechanism of basal ganglia diseases can be studied. It also provides a significant potential for diagnosis and treatment of basal ganglia diseases, including Parkinson's disease.